XIAO (LESTER) YU - Lester`s System Logs

XIAO (LESTER) YU
Department of Computer Science, North Carolina State University Raleigh, NC 27695
(919) · 518 · 4431 [email protected]
EDUCATION
North Carolina State University
Ph.D. in Computer Science
Advisor: Dr. Guoliang Jin
Visiting Student at University of Illinois at Urbana-Champaign
Host: Dr. Tao Xie
2011 - 2017 (Expected)
2013 - 2014
East China Normal University, Shanghai, China
M.S. in Computer Software & Theory
Advisor: Dr. Geguang Pu
B.S. in Software Engineering
2008 - 2011
2004 - 2008
RESEARCH INTERESTS
· Software Reliability and Performance
· Programming Languages
· Systems
· Automated Software Engineering
RESEARCH EXPERIENCES
Monitoring Cloud Infrastructures via Logs
Submitted to USENIX ATC 2015
2014 - 2015
In collaboration with NEC Laboratories America
· Proposed a log-based non-intrusive and lightweight approach to monitor task executions (e.g., starting
a virtual machine) for potential problems in cloud infrastructures.
· Addressed the challenges that log sequences from parallel task executions are usually interleaved, and
problematic executions may not have error log messages (e.g., performance degradation).
· The proposed approach can attribute each log message to a possible log sequence, and incrementally
check whether each log sequence with the additional message still conforms with the correct execution.
· Developed a failure-injection tool in OpenStack to systematically evaluate the effectiveness of the proposed approach.
Performance Analysis of Database Applications
In progress
2013 - Present
· Motivated by the wide adoption of databases in contemporary applications, investigated how applicationdatabase interactions can cause adverse performance impacts by studying and characterizing performance bugs in popular open-source projects.
· Based on the empirical evidence and challenges found by the study, planned to propose corresponding
solutions to detect and fix performance problems in database applications.
Performance Analysis of Device Drivers on Real-World Execution Traces
2012 - 2013
ASPLOS 2014
In collaboration with Microsoft Research Asia
· From large-scale real-world execution traces, identified that the system performance can be compromised by cost propagation, which is an adverse effect caused by contentions on multiple locks combining
with hierarchical dependencies among device drivers and the Windows kernel.
· To address multiple challenges posed by cost propagation on performance analysis, proposed a practical
two-step approach with effective data and pattern abstractions to measure performance impacts manifested through cost propagation, and to discover behavioral patterns closely related to performance
problems via contrast data mining.
Dynamic Test Generation for C Language
TASE 2009, QISC 2009, ENTCS 2011
2007 - 2011
· With improving the efficiency and the usability of symbolic execution, implemented the automated
testing tool CAUT 1 that aims at applying the technique to the industrial software testing.
· In order to facilitate the application of automated testing in industrial software, introduced a novel
program model and several particular algorithms as the supplement of the numerical constraint solver
to improve the approach of the test data generation for complex data structures and pointers.
· In order to relief the path explosion problem when testing compositional units, adopted side effect
analysis to eliminate the records of variables that are irrelevant to path conditions and partial execution
which can reduce the number of path exploring iterations.
· Designed parallel algorithms for dynamic test generation based on the partial order relations in the
sequential exploration of program paths.
Validation of Design Patterns in Object-Oriented Programs
IASTED 2010, AIMSEC 2011
2009 - 2010
· With selecting object-oriented concepts to describe structures and behaviors of programs with design
patterns, designed a relational calculus, and abstracted a specific object model in order to support the
validation algorithms.
· Established a methodology and corresponding tool to detect and validate the usage of design patterns
in Java program using the relational calculus and the abstract object model.
Analysis and Verification of rCOS
ENTCS 2008, APSEC 2008
2
2007 - 2008
· Extended the original version of rCOS to support assertions, invariants and parallelism, and proposed
an approach to transform the extended rCOS model to SPIN model in order to provide a straight way
for model verification.
· Constructed the operational semantic model for rCOS including parallelism, which clarified the real
effects of rCOS programs, and provided a secure guide for the implementation of rCOS.
PUBLICATIONS
1. Xiao Yu, Pallavi Joshi, Jianwu Xu, Guoliang Jin, Hui Zhang, and Guofei Jiang. CloudSeer: Monitoring Cloud Infrastructures via Interleaved Logs. Submitted to 2015 USENIX Annual Technical
Conference.
2. Xiao Yu, Shi Han, Dongmei Zhang, and Tao Xie. Comprehending Performance from Real-World
Execution Traces: A Device-Driver Case. In Proceedings of the 19th International Conference on
Architectural Support for Programming Languages and Operating Systems (ASPLOS), 2014.
3. Wei Yang, Xiao Yu. AT-EASE: A Tool for Early and Quick Usability Evaluation of Smartphone
Applications. Technical Report, 2013.
4. Xiao Yu, Shuai Sun, Geguang Pu, Siyuan Jiang, and Zheng Wang. A Parallel Approach to Concolic
Testing with Low-cost Synchronization. Electronic Notes in Theoretical Computer Science (ENTCS),
Volume 274 (August 2011), pages 83-96, 2011.
5. Kang Miao, Xiao Yu, Siyuan Jiang, and Ji Zhao, Run-time Discovery of Java Design Patterns. In Proceedings of the 2011 Second International Conference on Artificial Intelligence, Management Science
and Electronic Commerce (AIMSEC), pages 3329-3332, 2011.
1
2
CAUT(C Analysis and Unit Testing): http://lab205.org/caut
rCOS: http://rcos.iist.unu.edu/
6. Kang Miao, Xiao Yu, Ji Zhao, and Yu Shen. Java Design Pattern Recognition Based on Relational
Calculus. Application Research of Computers (in Chinese), Volume 27, Issue 9, pages 3425-3430,
2010.
7. Libo Feng, Xiao Yu, Geguang Pu, Huibiao Zhu, and Bin Gu. Property Checking for Design Patterns.
In Proceedings of the IASTED International Conference on Software Engineering, pages 87-94, 2010.
8. Tao Sun, Zheng Wang, Geguang Pu, Xiao Yu, Zongyan Qiu, and Bin Gu. Towards Scalable Compositional Test Generation. In Proceedings of the 2009 Ninth International Conference on Quality
Software (QISC), pages 353-358, 2009.
9. Zheng Wang, Xiao Yu, Tao Sun, Geguang Pu, Zuohua Ding, and JueLiang Hu. Test Data Generation
for Derived Types in C Program. In Proceedings of the 2009 Third IEEE International Symposium
on Theoretical Aspects of Software Engineering (TASE), pages 155- 162, 2009.
10. Zheng Wang, Xiao Yu, Geguang Pu, Libo Feng, Huibiao Zhu, and Jifeng He. Execution Semantics
for rCOS. In Proceedings of the 2008 15th Asia-Pacific Software Engineering Conference (APSEC),
pages 119-126, 2008.
11. Xiao Yu, Zheng Wang, Geguang Pu, Dingding Mao, and Jing Liu. The Verification of rCOS Using
Spin. Electronic Notes in Theoretical Computer Science (ENTCS), Volume 207 (April 2008), pages
49-67, 2008.
TEACHING EXPERIENCES
Teaching Assistants
· CSC 568: Enterprise Storage Architecture
Lecturer: Dr. Vincent W. Freeh
· CSC 506: Architecture of Parallel Computers
Lecturer: Dr. Edward F. Gehringer
· CSC 510: Software Engineering
Lecturer: Dr. Annie I. Ant´
on
Spring 2015
Spring 2012, Spring 2015
Fall 2011
PROFESSIONAL ACTIVITIES
Subreviewers
Student Volunteer
ASE 2014; ICSM 2012; ICST 2014; ISSTA 2012, 2013; MSR 2014; OOPSLA 2013
FSE 2012